Skip to content

Commit

Permalink
update omero_hyperstack_to_gastruloid_measurement + add README with C…
Browse files Browse the repository at this point in the history 
…HANGELOG
  • Loading branch information
@lldelisle
lldelisle committed Dec 20, 2023
1 parent 4ac9b1d commit dbadf64
Show file tree
Hide file tree
Showing 5 changed files with 202 additions and 133 deletions.
2 changes: 0 additions & 2 deletions .tt_skip
View file
Original file line number Diff line number Diff line change
@@ -1,6 +1,4 @@
tools/measure_gastruloids
tools/incucyte_stack_and_upload_omero/
tools/hyperstack_to_bleach_corrected_movie
tools/omero_hyperstack_to_fluo_measurements_on_gastruloid
tools/omero_hyperstack_to_gastruloid_measurements
tools/omero_clean_rois_tables
2 changes: 1 addition & 1 deletion tools/omero_hyperstack_to_gastruloid_measurements/.shed.yml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,5 +3,5 @@ categories:
description: Analyse Hyperstack on OMERO server to segment gastruloid and compute measurements
name: omero_hyperstack_to_gastruloid_measurements
owner: lldelisle
long_description: Uses a groovy to get images from OMERO, use an ilastik project to get propability and generate a mask. The potential gastruloids will be indentified by analyze Particles. On each ROI, the elongation index will be computed.
long_description: Uses a groovy to get images from OMERO, use an ilastik project to get propability and generate a mask or simply autothreshold to mask. The potential gastruloids will be indentified by analyze Particles. On each ROI, the elongation index will be computed.
remote_repository_url: https://github.com/lldelisle/tools-lldelisle/tree/master/tools/omero_hyperstack_to_gastruloid_measurements
207 changes: 112 additions & 95 deletions ...yperstack_to_gastruloid_measurements/1-omero_timelapse_image_to_measurements_phase.groovy
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@
// merge the analysis script with templates available at
// https://github.com/BIOP/OMERO-scripts/tree/main/Fiji

// Last modification: 2023-07-28
// Last modification: 2023-12-20

/*
* = COPYRIGHT =
Expand All @@ -29,7 +29,8 @@
* ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/

// This macro will use ilastik to detect ROIs
// This macro will use ilastik or convert to mask
// to detect ROIs
// measure and compute elongation index
// It may also regenerate a ROI of background

Expand Down Expand Up @@ -88,6 +89,7 @@ import ij.plugin.Duplicator
import ij.plugin.frame.RoiManager
import ij.plugin.HyperStackConverter
import ij.plugin.ImageCalculator
import ij.plugin.Thresholder
import ij.Prefs
import ij.process.FloatPolygon
import ij.process.ImageProcessor
Expand Down Expand Up @@ -402,7 +404,7 @@ def processDataset(Client user_client, DatasetWrapper dataset_wrp,
Boolean headless_mode, Boolean debug, String tool_version,
Boolean use_existing, String final_object, Boolean rescue,
Integer ilastik_label_BG, Double probability_threshold_BG,
Boolean keep_only_largest) {
Boolean keep_only_largest, String segmentation_method) {
robustlyGetAll(dataset_wrp, "image", user_client).each{ ImageWrapper img_wrp ->
processImage(user_client, img_wrp,
ilastik_project, ilastik_project_type,
Expand All @@ -414,7 +416,7 @@ def processDataset(Client user_client, DatasetWrapper dataset_wrp,
headless_mode, debug, tool_version,
use_existing, final_object, rescue,
ilastik_label_BG, probability_threshold_BG,
keep_only_largest)
keep_only_largest, segmentation_method)
}
}

Expand All @@ -429,7 +431,7 @@ def processSinglePlate(Client user_client, PlateWrapper plate_wrp,
Boolean headless_mode, Boolean debug, String tool_version, Boolean use_existing,
String final_object, Boolean rescue,
Integer ilastik_label_BG, Double probability_threshold_BG,
Boolean keep_only_largest) {
Boolean keep_only_largest, String segmentation_method) {
robustlyGetAll(plate_wrp, "well", user_client).each{ well_wrp ->
processSingleWell(user_client, well_wrp,
ilastik_project, ilastik_project_type,
Expand All @@ -441,7 +443,7 @@ def processSinglePlate(Client user_client, PlateWrapper plate_wrp,
headless_mode, debug, tool_version,
use_existing, final_object, rescue,
ilastik_label_BG, probability_threshold_BG,
keep_only_largest)
keep_only_largest, segmentation_method)
}
}

Expand All @@ -456,7 +458,7 @@ def processSingleWell(Client user_client, WellWrapper well_wrp,
Boolean headless_mode, Boolean debug, String tool_version, Boolean use_existing,
String final_object, Boolean rescue,
Integer ilastik_label_BG, Double probability_threshold_BG,
Boolean keep_only_largest) {
Boolean keep_only_largest, String segmentation_method) {
well_wrp.getWellSamples().each{
processImage(user_client, it.getImage(),
ilastik_project, ilastik_project_type,
Expand All @@ -468,7 +470,7 @@ def processSingleWell(Client user_client, WellWrapper well_wrp,
headless_mode, debug, tool_version,
use_existing, final_object, rescue,
ilastik_label_BG, probability_threshold_BG,
keep_only_largest)
keep_only_largest, segmentation_method)
}
}

Expand All @@ -483,7 +485,7 @@ def processImage(Client user_client, ImageWrapper image_wrp,
Boolean headless_mode, Boolean debug, String tool_version,
Boolean use_existing, String final_object, Boolean rescue,
Integer ilastik_label_BG, Double probability_threshold_BG,
Boolean keep_only_largest) {
Boolean keep_only_largest, String segmentation_method) {

IJ.run("Close All", "")
IJ.run("Clear Results")
Expand Down Expand Up @@ -596,96 +598,105 @@ def processImage(Client user_client, ImageWrapper image_wrp,
}
if (!use_existing) {
// We compute the segmentation
File output_path = new File (output_directory, image_basename+"_ilastik_" + ilastik_project_short_name + "_output.tif" )
ImagePlus predictions_imp
FileSaver fs
if(output_path.exists()) {
println "USING EXISTING ILASTIK OUTPUT"
predictions_imp = IJ.openImage( output_path.toString() )
} else {
/**
* ilastik
*/
println "Starting ilastik"

// get ilastik predictions for each time point of the Time-lapse but all at the same time
ImagePlus ilastik_input_original = new Duplicator().run(imp, ilastik_input_ch, ilastik_input_ch, 1, 1, 1, nT);

ImagePlus gb_imp = ilastik_input_original.duplicate()
IJ.run(gb_imp, "Gaussian Blur...", "sigma=100 stack")
ImagePlus ilastik_input = ImageCalculator.run(ilastik_input_original, gb_imp, "Divide create 32-bit stack")
if (!headless_mode) {ilastik_input.show()}
// can't work without displaying image
// IJ.run("Run Pixel Classification Prediction", "projectfilename="+ilastik_project+" inputimage="+ilastik_input.getTitle()+" pixelclassificationtype=Probabilities");
//
// to use in headless_mode more we need to use a commandservice
def predictions_imgPlus
if (ilastik_project_type == "Regular") {
predictions_imgPlus = cmds.run( IlastikPixelClassificationCommand.class, false,
'inputImage', ilastik_input,
'projectFileName', ilastik_project,
'pixelClassificationType', "Probabilities").get().getOutput("predictions")
if (segmentation_method == "ilastik") {
File output_path = new File (output_directory, image_basename+"_ilastik_" + ilastik_project_short_name + "_output.tif" )
ImagePlus predictions_imp
FileSaver fs
if(output_path.exists()) {
println "USING EXISTING ILASTIK OUTPUT"
predictions_imp = IJ.openImage( output_path.toString() )
} else {
predictions_imgPlus = cmds.run( IlastikAutoContextCommand.class, false,
'inputImage', ilastik_input,
'projectFileName', ilastik_project,
'AutocontextPredictionType', "Probabilities").get().getOutput("predictions")
}
// to convert the result to ImagePlus : https://gist.github.com/GenevieveBuckley/460d0abc7c1b13eee983187b955330ba
predictions_imp = ImageJFunctions.wrap(predictions_imgPlus, "predictions")
/**
* ilastik
*/
println "Starting ilastik"

// get ilastik predictions for each time point of the Time-lapse but all at the same time
ImagePlus ilastik_input_original = new Duplicator().run(imp, ilastik_input_ch, ilastik_input_ch, 1, 1, 1, nT);

ImagePlus gb_imp = ilastik_input_original.duplicate()
IJ.run(gb_imp, "Gaussian Blur...", "sigma=100 stack")
ImagePlus ilastik_input = ImageCalculator.run(ilastik_input_original, gb_imp, "Divide create 32-bit stack")
if (!headless_mode) {ilastik_input.show()}
// can't work without displaying image
// IJ.run("Run Pixel Classification Prediction", "projectfilename="+ilastik_project+" inputimage="+ilastik_input.getTitle()+" pixelclassificationtype=Probabilities");
//
// to use in headless_mode more we need to use a commandservice
def predictions_imgPlus
if (ilastik_project_type == "Regular") {
predictions_imgPlus = cmds.run( IlastikPixelClassificationCommand.class, false,
'inputImage', ilastik_input,
'projectFileName', ilastik_project,
'pixelClassificationType', "Probabilities").get().getOutput("predictions")
} else {
predictions_imgPlus = cmds.run( IlastikAutoContextCommand.class, false,
'inputImage', ilastik_input,
'projectFileName', ilastik_project,
'AutocontextPredictionType', "Probabilities").get().getOutput("predictions")
}
// to convert the result to ImagePlus : https://gist.github.com/GenevieveBuckley/460d0abc7c1b13eee983187b955330ba
predictions_imp = ImageJFunctions.wrap(predictions_imgPlus, "predictions")

predictions_imp.setTitle("ilastik_output")
predictions_imp.setTitle("ilastik_output")

// save file
fs = new FileSaver(predictions_imp)
fs.saveAsTiff(output_path.toString() )
}
if (!headless_mode) { predictions_imp.show() }
// save file
fs = new FileSaver(predictions_imp)
fs.saveAsTiff(output_path.toString() )
}
if (!headless_mode) { predictions_imp.show() }

/**
* From the "ilastik predictions of the Time-lapse" do segmentation and cleaning
*/
/**
* From the "ilastik predictions of the Time-lapse" do segmentation and cleaning
*/

// Get a stack of ROI for background:
if (ilastik_label_BG != 0) {
ImagePlus mask_imp_BG = new Duplicator().run(predictions_imp, ilastik_label_BG, ilastik_label_BG, 1, 1, 1, nT)
// Apply threshold:
IJ.setThreshold(mask_imp_BG, probability_threshold_BG, 100.0000)
Prefs.blackBackground = true
IJ.run(mask_imp_BG, "Convert to Mask", "method=Default background=Dark black")
if (!headless_mode) { mask_imp_BG.show() }
IJ.run(mask_imp_BG, "Analyze Particles...", "stack show=Overlay")
Overlay ov_BG = mask_imp_BG.getOverlay()
Overlay ov_BG_Combined = new Overlay()
for (int t=1;t<=nT;t++) {
// Don't ask me why we need to refer to Z pos and not T/Frame
ArrayList<Roi> all_rois_inT = ov_BG.findAll{ roi -> roi.getZPosition() == t}
println "There are " + all_rois_inT.size() + " in time " + t
if (all_rois_inT.size() > 0) {
ShapeRoi current_roi = new ShapeRoi(all_rois_inT[0] as Roi)
for (i = 1; i < all_rois_inT.size(); i++) {
current_roi = current_roi.or(new ShapeRoi(all_rois_inT[i] as Roi))
// Get a stack of ROI for background:
if (ilastik_label_BG != 0) {
ImagePlus mask_imp_BG = new Duplicator().run(predictions_imp, ilastik_label_BG, ilastik_label_BG, 1, 1, 1, nT)
// Apply threshold:
IJ.setThreshold(mask_imp_BG, probability_threshold_BG, 100.0000)
Prefs.blackBackground = true
IJ.run(mask_imp_BG, "Convert to Mask", "method=Default background=Dark black")
if (!headless_mode) { mask_imp_BG.show() }
IJ.run(mask_imp_BG, "Analyze Particles...", "stack show=Overlay")
Overlay ov_BG = mask_imp_BG.getOverlay()
Overlay ov_BG_Combined = new Overlay()
for (int t=1;t<=nT;t++) {
// Don't ask me why we need to refer to Z pos and not T/Frame
ArrayList<Roi> all_rois_inT = ov_BG.findAll{ roi -> roi.getZPosition() == t}
println "There are " + all_rois_inT.size() + " in time " + t
if (all_rois_inT.size() > 0) {
ShapeRoi current_roi = new ShapeRoi(all_rois_inT[0] as Roi)
for (i = 1; i < all_rois_inT.size(); i++) {
current_roi = current_roi.or(new ShapeRoi(all_rois_inT[i] as Roi))
}
// Update the position before adding to the ov_BG_Combined
current_roi.setPosition( ilastik_input_ch, 1, t)
current_roi.setName("Background_t" + t)
ov_BG_Combined.add(current_roi)
}
// Update the position before adding to the ov_BG_Combined
current_roi.setPosition( ilastik_input_ch, 1, t)
current_roi.setName("Background_t" + t)
ov_BG_Combined.add(current_roi)
}
IJ.run("Clear Results")
println "Store " + ov_BG_Combined.size() + " BG ROIs on OMERO"
// Save ROIs to omero
robustlysaveROIs(image_wrp, user_client, ROIWrapper.fromImageJ(ov_BG_Combined as List))
}
IJ.run("Clear Results")
println "Store " + ov_BG_Combined.size() + " BG ROIs on OMERO"
// Save ROIs to omero
robustlysaveROIs(image_wrp, user_client, ROIWrapper.fromImageJ(ov_BG_Combined as List))
}

// Get only the channel for the gastruloid/background prediction
mask_imp = new Duplicator().run(predictions_imp, ilastik_label_OI, ilastik_label_OI, 1, 1, 1, nT);
// Get only the channel for the gastruloid/background prediction
mask_imp = new Duplicator().run(predictions_imp, ilastik_label_OI, ilastik_label_OI, 1, 1, 1, nT);

// Apply threshold:
IJ.setThreshold(mask_imp, probability_threshold, 100.0000);
Prefs.blackBackground = true;
IJ.run(mask_imp, "Convert to Mask", "method=Default background=Dark black");

} else {
// Get only the channel with bright field
mask_imp = new Duplicator().run(imp, ilastik_input_ch, ilastik_input_ch, 1, 1, 1, nT);
// Run convert to mask
(new Thresholder()).convertStackToBinary(mask_imp);
}
// This title will appear in the result table
mask_imp.setTitle(image_basename)
// Apply threshold:
IJ.setThreshold(mask_imp, probability_threshold, 100.0000);
Prefs.blackBackground = true;
IJ.run(mask_imp, "Convert to Mask", "method=Default background=Dark black");
if (!headless_mode) { mask_imp.show() }

// clean the mask a bit
Expand Down Expand Up @@ -775,8 +786,13 @@ def processImage(Client user_client, ImageWrapper image_wrp,
rt.setValue("Unit", row, scale_unit)
rt.setValue("Date", row, now)
rt.setValue("Version", row, tool_version)
rt.setValue("IlastikProject", row, ilastik_project_short_name)
rt.setValue("ProbabilityThreshold", row, probability_threshold)
if (segmentation_method == "ilastik") {
rt.setValue("IlastikProject", row, ilastik_project_short_name)
rt.setValue("ProbabilityThreshold", row, probability_threshold)
} else {
rt.setValue("IlastikProject", row, "NA")
rt.setValue("ProbabilityThreshold", row, "NA")
}
rt.setValue("MinSizeParticle", row, min_size_particle)
rt.setValue("MinDiameter", row, minimum_diameter)
rt.setValue("ClosenessTolerance", row, closeness_tolerance)
Expand Down Expand Up @@ -1080,7 +1096,7 @@ def processImage(Client user_client, ImageWrapper image_wrp,
// In simple-omero-client
// Strings that can be converted to double are stored in double
// In order to build the super_table, tool_version should stay String
String tool_version = "Phase_v20230728"
String tool_version = "White_v20231220"

// User set variables

Expand All @@ -1095,6 +1111,7 @@ String tool_version = "Phase_v20230728"

#@ String(visibility=MESSAGE, value="Parameters for segmentation/ROI", required=false) msg2
#@ Boolean(label="Use existing segmentation (values below in the section will be ignored)") use_existing
#@ String(label="Segmentation Method", choices={"convert_to_mask","ilastik"}) segmentation_method
#@ Boolean(label="<html>Run in rescue mode<br/>(only segment images without tables)</html>", value=false) rescue
#@ File(label="Ilastik project") ilastik_project
#@ String(label="Ilastik project short name") ilastik_project_short_name
Expand Down Expand Up @@ -1206,7 +1223,7 @@ if (user_client.isConnected()) {
headless_mode, debug, tool_version,
use_existing, "image", rescue,
ilastik_label_BG, probability_threshold_BG,
keep_only_largest)
keep_only_largest, segmentation_method)
break
case "dataset":
DatasetWrapper dataset_wrp = robustlyGetOne(id, "dataset", user_client)
Expand All @@ -1233,7 +1250,7 @@ if (user_client.isConnected()) {
headless_mode, debug, tool_version,
use_existing, "dataset", rescue,
ilastik_label_BG, probability_threshold_BG,
keep_only_largest)
keep_only_largest, segmentation_method)
// upload the table on OMERO
super_table.setName(table_name + "_global")
robustlyAddAndReplaceTable(dataset_wrp, user_client, super_table)
Expand Down Expand Up @@ -1263,7 +1280,7 @@ if (user_client.isConnected()) {
headless_mode, debug, tool_version,
use_existing, "well", rescue,
ilastik_label_BG, probability_threshold_BG,
keep_only_largest)
keep_only_largest, segmentation_method)
// upload the table on OMERO
super_table.setName(table_name + "_global")
robustlyAddAndReplaceTable(well_wrp, user_client, super_table)
Expand Down Expand Up @@ -1293,7 +1310,7 @@ if (user_client.isConnected()) {
headless_mode, debug, tool_version,
use_existing, "plate", rescue,
ilastik_label_BG, probability_threshold_BG,
keep_only_largest)
keep_only_largest, segmentation_method)
// upload the table on OMERO
super_table.setName(table_name + "_global")
robustlyAddAndReplaceTable(plate_wrp, user_client, super_table)
Expand Down
Loading

0 comments on commit dbadf64

Please sign in to comment.